Over the past few weeks, we’ve had several PAADs that discussed pediatric preoperative fasting guidelines and the growing awareness of our need to rethink them. This really is not a new topic and has been debated for at least 20 years. I asked Mark Schreiner, a frequent PAAD contributor and close colleague and friend of Ron Litman, to discuss his pivotal role in studying this issue and the story behind the story of how he went about doing it. Myron Yaster MD
It may be hard to believe now but we used to starve children for hours and hours before surgery. Even though clear liquids empty from the stomach rapidly everyone maintained an npo after midnight regimen for all patients, including those having afternoon surgery. Roger Maltby started to challenge the dogma of npo after midnight; reading his papers made total sense to me. As it happened my first faculty office was situated at the end of a long corridor containing the day surgery pre-discharge unit. As I watched children leaving day surgery for home, it was not infrequent that I witnessed them regurgitate the clear liquids that they had just consumed.
As a pediatric resident, vomiting and diarrhea complaints were a frequent cause for ER visits. We were taught to have the parents wait at least 2 hours before resuming attempts to administer clear fluids. And yet nauseated post-op children were required to drink before they were allowed to leave the hospital. It struck me that we were doing things entirely backwards; we were starving children pre-op and forcing nauseated children to drink post-op.
I designed a pair of clinical trials (my first), to see if my hunches were correct or not.
Schreiner MS, Triebwasser A, Keon TP. Ingestion of liquids compared with preoperative fasting in pediatric outpatients. Anesthesiology. 1990 Apr;72(4):593-7. PMID: 2321772.
We randomized 121 children to either follow the usual fasting guidelines or a liberalized guideline which permitted unlimited volumes of clear fluids (with the exception of the last drink which was limited to 8 fluid ounces) up until 2 hours before surgical start time. Previous studies by Bill Splinter and others had limited the volume of fluid to no more than 3 mL/kg. Suffice it to say that there were no differences in gastric fluid volume or pH between the two groups. The parents of children assigned to the liberalized feeding instructions found it easier to comply with the instructions and thought their children better tolerated the preoperative experience. Parents of children following the usual fasting guideline, assessed their children as being more irritable if fasted after midnight.
At the SPA meeting the following spring there was a session on controversial topics where the attendees were polled. As I recall, over 90% of attendees reported that their institution had already adopted liberalized fasting guideline. In under a year, this issue had rapidly become not very controversial which was incredibly gratifying.
Schreiner MS, Nicolson SC, Martin T, Whitney L. Should children drink before discharge from day surgery? Anesthesiology. 1992 Apr;76(4):528-33. PMID: 1550277.
With part one of the plan completed, I then conducted a second clinical trial to determine whether mandatory drinking was really necessary to ensure that children could be discharged home after day surgery. The fear was that if a child didn’t drink in hospital, they might not at home with resulting dehydration. We randomized almost 1000 children to either the traditional discharge instructions, which meant that had to drink and tolerate clear liquids in order to go home or to elective drinking. This meant that they could be offered clear liquids but there was no requirement to drink them. Since we couldn’t have children in adjoining beds with different discharge instructions and to avoid confusion amongst the nurses, the groups were randomized by week. The elective drinkers vomited less in the hospital (14% vs 23%) and overall (32% vs 39%) than the mandatory drinkers. No children from either group required readmission for dehydration or intractable vomiting.
At the time I conducted these studies, I had no idea what a power analysis was. All I knew was that larger sample sizes were better than smaller ones. As a clinical trialist I was pretty green. Rereading the first paper now, I can see a number of amateur mistakes. (1) Randomization Imbalance. There were 53 children in drinking group and 68 in the usual care group. Some of that was due to cancellations but mostly, it was due to using simple randomization instead of block or permuted block randomization which would have ensured that the two groups were closer in size. (2) Presenting P values in the Demographic Table (Table 1). Randomization is intended to distribute factors known and unknown equally between the two groups. Hypothesis testing is not appropriate for comparing those groups. I would think that most journals no longer allow inclusion of P values any longer. (3) Analysis. The analysis focused on superiority even though that wasn’t really the objective. Since we had no expectation or rationale for pre-operative drinking to be better than fasting, we should have determined whether or not it was non-inferior to the standard rather than superior.
As a result of these two studies (and a few other related studies), I was invited to serve on two of the earliest ASA Practice Guideline committees, one on preoperative fasting and one on postanesthetic care. Even though we were cooped up in a Chicago airport hotel room for several days at a time, getting the chance to meet and interact with clinical experts and statisticians versed in meta-analysis coming from all over the country was pretty exciting. Mark Schreiner MD